This letter proposes a novel learned feasibility estimator that considers multi-modal grasp poses for grasp and motion planning. Grasp poses inherently have multi-modal structures, that is, continuous and discrete parameters. Mixed-integer programming (MIP) is one method that solves these multi-modal problems. However, searching for all the discrete parameters costs considerable time. Therefore, by learning the feasibility of each mode from the geometric variables, the problem can be solved efficiently within a given time limit. The feasibility of grasp poses is related to the pose of the object and nearby obstacles. Utilizing this information, we introduce learned geometric feasibility (LGF), which prioritizes the integer search of MIP. LGF is scalable to multiple robots and environments because it learns the feasibility using object-oriented information. It has been demonstrated to improve the number of solved MIP problems within the time limit and to be applicable to various environmental settings.

This paper presents a novel framework for the complete assembly of a chair based on assembly instructions. First, the framework utilizes task templates and task compiler that construct robot tasks from the given assembly instructions such as connection relationships and assembly sequences. Each task template contains the motion planning and control strategies to accomplish each task. The task compiler converts assembly instructions into a series of robotic tasks by utilizing the task templates. Second, the framework proposes enhanced techniques related to the motion planning and control. In the motion planning, optimized goal configurations are computed away from the joint limit and singularity to successfully execute the subsequent assembly motion. Regarding control strategy, macro- and micro-motions are introduced to modularize motions covering various assembly types in the presence of uncertainties. To validate the framework, the assembly of a chair was performed completely. Additionally, the most used task, pin insertion task for both simple and complex shapes, was selected as a benchmark task and evaluated by the success rate. 

This paper proposes a novel inverse kinematics (IK) solver of articulated robotic systems for path planning. IK is a traditional but essential problem for robot manipulation. Recently, data-driven methods have been proposed to quickly solve the IK for path planning. These methods can handle a large amount of IK requests at once with the advantage of GPUs. However, the accuracy is still low, and the model requires considerable time for training. Therefore, we propose an IK solver that improves accuracy and memory efficiency by utilizing the continuous hidden dynamics of Neural ODE. The performance is compared using multiple robots.

In this study, a new method is proposed to generate rigid grasp candidates for assembly tasks. Although the majority of the recent research is aimed at the generation of candidates for grasping several types of objects, our method focuses on tasks requiring precise and rigid contact. For instance, assembling furniture requires higher grasping force and precision than bin picking tasks or pick and place tasks. This means that the pose information of the object should be accurate for assembly tasks, and the object should not slip while being held. Given these constraints, we generated possible grasp pose candidates. The proposed algorithm is verified by performing real robot experiments in which furniture parts are assembled.